This study investigated whether task instruction affected comprehension of multiple conflicting-view texts after controlling for a number of individual difference variables and whether the effects of task instruction varied as a result of post-reading assessment tasks. Recruited from a First-Year Composition course, 64 participants received a task instruction that set the goal of seeking an alternative explanation and reframed argument as a process of conversation and exploration. Multiple-text comprehension was measured by a synthesis writing task and an argument writing task. The results showed that participants given the instruction outperformed participants in the control condition, as reflected in their argument writing. Among the control variables, beliefs about argumentation, need for cognition and topic interest were positively correlated with comprehension measures. These results highlight the importance of cultivating divergent thinking of multiple perspectives rather than dichotomous thinking of pros and cons. Pedagogical implications are discussed.